1
0
mirror of https://github.com/danog/PHP-Parser.git synced 2024-11-30 04:19:30 +01:00
Commit Graph

416 Commits

Author SHA1 Message Date
Nikita Popov
d287c167bc Pretty print: Handle Error in ClassConstFetch 2017-02-03 21:53:02 +01:00
Nikita Popov
79afd56565 Add NodeFinder class
To simplify basic node finding operations.
2017-01-29 23:20:53 +01:00
Nikita Popov
5cc2750ebc Merge branch '3.x'
Conflicts:
	lib/PhpParser/NodeVisitor.php
2017-01-29 22:36:33 +01:00
Nikita Popov
58e7881e98 Implement NodeTraverser::STOP_TRAVERSAL
Conflicts:
	lib/PhpParser/NodeVisitor.php
2017-01-29 22:35:40 +01:00
Nikita Popov
e072c789d1 Merge branch '3.x'
Conflicts:
	test/PhpParser/PrettyPrinterTest.php
2017-01-29 21:57:37 +01:00
Nikita Popov
fd7ac25108 Throw if pretty-printing Error node 2017-01-29 21:56:21 +01:00
Nikita Popov
bfea338d36 Update doc comments after previous comment
Make some of the type annotations more accurate, and complete the
generated doc-comments to be complete (with description and
parameter annotations.)
2017-01-26 00:16:54 +01:00
Nikita Popov
cb5dd28985 Don't include whitespace directly in catch/finally print 2017-01-21 21:25:48 +01:00
Nikita Popov
5e565e8046 Support insertion of nullable nodes
Still incomplete in some places and the formatting is not always
ideal.
2017-01-21 21:20:42 +01:00
Nikita Popov
b9b6aeeed9 Support format-preserving node removal
Take care of stripping surrouding tokens appropriately.
2017-01-21 17:52:50 +01:00
Nikita Popov
0607450f78 Add FinderVisitor
Allows finding nodes based on a filter callback.
2017-01-21 15:41:24 +01:00
Nikita Popov
b5fb6f2d0a Don't require useNopStatements=false for format preservation
Instead assign attributes on Nop nodes and in the pretty printer
specially handle end<start offsets. It's a somewhat weird case,
but not wrong per se given the meaning the offsets have.
2017-01-20 22:27:51 +01:00
Nikita Popov
48d3243abe Drop last vestiges of consistent var mode flag 2017-01-19 23:46:25 +01:00
Nikita Popov
d7f3c4f9d3 Renamve Param::$name to Param::$var
As it now contains a Variable node.
2017-01-19 23:39:15 +01:00
Nikita Popov
a79306ccd9 Rename StaticVar::$name to $var
As it now holds a Variable
2017-01-19 23:35:31 +01:00
Nikita Popov
6238f5f9f9 Adjust code for constitent var mode 2017-01-19 23:32:49 +01:00
Nikita Popov
67274b9594 Enforce useConsistentVariableNames 2017-01-19 23:24:43 +01:00
Nikita Popov
7623d20f69 Automatically wrap in Stmt\Expression in builders 2017-01-19 22:39:21 +01:00
Nikita Popov
33552764ad Perform manual test updates 2017-01-19 22:31:29 +01:00
Nikita Popov
953f8c9631 Perform automated test update 2017-01-19 22:25:22 +01:00
Nikita Popov
b0c962911e Add test porting infrastructure 2017-01-19 22:24:48 +01:00
Nikita Popov
1bfbd7bcc8 Make useExpressionStatement only supported mode 2017-01-19 21:15:26 +01:00
Nikita Popov
065c720c28 Merge branch 'formatPreservingPrint' 2017-01-19 21:06:32 +01:00
Nikita Popov
0c9c8d58ab Drop deprecated $type subnodes
These have been replaced by $flags in 3.0
2017-01-19 21:00:44 +01:00
Nikita Popov
f21309f52f Add NameResolver mode that does not modify nodes 2016-12-26 21:45:33 +01:00
Nikita Popov
4d2a4d02b0 Add first shot at format preserving pretty printer 2016-12-26 18:28:49 +01:00
Nikita Popov
9b2d35d1ac Add expression statement mode 2016-12-24 23:54:24 +01:00
Nikita Popov
0f582e1708 Add VarLikeIdentifier
For representing Identifiers that have an implicit leading $.

With this done, maybe go one step further?
 * Rename VarLikeIdentifier -> VarIdentifier / VarName
 * Use VarIdentifier / VarName also as an inner node in Variable.
   Not sure if this adds any real value.
2016-12-23 12:39:27 +01:00
Nikita Popov
d32d937d47 Use Identifier for property names as well 2016-12-23 00:25:45 +01:00
Nikita Popov
f46c909b5a Represent prop in static prop fetch using Identifier 2016-12-23 00:22:30 +01:00
Nikita Popov
a947e731c3 Add useConsistentVariableNodes mode
The parameter case is a bit weird, because the subnode is called
"name" here, rather than "var". Nothing we can do about that in
this version though.

The two parser options might be merged. I've kept it separate,
because I think this variable representation should become the
default (or even only representation) in the future, while I'm
less sure about the Identifier thing.
2016-12-23 00:10:59 +01:00
Nikita Popov
122f449960 Represent builtin types using Identifier as well 2016-12-22 22:23:30 +01:00
Nikita Popov
6bcc6c31dd Add useIdentifierNodes mode to parser
In this mode non-namespaced names that are currently represented
using strings will be represented using Identifier nodes instead.
Identifier nodes have a string $name subnode and coerce to string.

This allows preserving attributes and in particular location
information on identifiers.
2016-12-22 21:15:44 +01:00
Felix Becker
3e8c8d248d Add originalName attribute in NameResolver
For now gated behind a preserveOriginalNames option.
2016-12-22 20:25:02 +01:00
Nikita Popov
58970e2a37 Improve LNumber/DNumber pretty printing
* Support PHP_INT_MIN
* Support negative binary/octal/hex numbers
* Support INF/-INF/NAN in namespaces
2016-12-11 16:31:59 +01:00
Nikita Popov
c1e0bab4f8 Add support for negative interpolated offsets
A late addition to PHP 7.1 which I missed.
2016-12-11 13:44:17 +01:00
Nikita Popov
70319e27ee Parse 0 in "$a[0]" as LNumber (#325) 2016-12-07 20:24:00 +01:00
Nikita Popov
5219f75719 Fix pretty-printing of nullable types 2016-12-06 12:26:21 +01:00
Ondrej Mirtes
a485ecd7ba NameResolver - resolve Name in NullableType 2016-12-06 12:21:30 +01:00
Nikita Popov
3e158a2313 Wrap List_ in ArrayItem
This was correctly done for the 'key'=>list() form, but not for
unkeyed nested lists.
2016-12-01 13:32:37 +01:00
Nikita Popov
bcdfb703d5 Cleanup imports
Thanks PhpStorm :)
2016-11-23 22:58:18 +01:00
Nikita Popov
27281e9130 Fix attribute assignment for Error nodes 2016-11-23 22:51:32 +01:00
Nikita Popov
098294beec Support !!positions parser test mode
And use it for the group use prefix position test that was
previously implemented as a separate test.
2016-11-23 22:36:48 +01:00
Nikita Popov
b02f8ac07d Add support for dumping positions in NodeDumper 2016-11-23 22:25:17 +01:00
Nikita Popov
e52ffc4447 Support recovery from Foo:: 2016-11-22 19:47:04 +01:00
Nikita Popov
c5cdd5ad73 Support recovery from free-standing $ 2016-11-21 17:01:39 +01:00
Nikita Popov
c0630f8169 Support recovery for new without class name 2016-11-21 16:51:53 +01:00
Nikita Popov
fa7357b483 Represent empty Name::slice() using null
Instead of a Name([]) dummy value, that is invalid in other
contexts.
2016-10-22 17:02:38 +02:00
Nikita Popov
91cb82d3d2 Explicitly support Name copy construction
It already worked beforehand by accident ... make clear it's
actually supported.
2016-10-22 16:41:58 +02:00
Nikita Popov
7672b974ff Remove Name::append() and Name::prepend() 2016-10-22 00:25:15 +02:00
Nikita Popov
f99a96e0a2 Introduce ErrorHandler
Add ErrorHandler interface, as well as ErrorHandler\Throwing
and ErrorHandler\Collecting. The error handler is passed to
Parser::parse(). This supersedes the throwOnError option.

NameResolver now accepts an ErrorHandler in the ctor.
2016-10-16 22:12:46 +02:00
Nikita Popov
90834bff8e Add namespacedName attribute on runtime-resolved names
The NameResolver now adds a namespacedName attribute on function/
const names which cannot be statically resolved.
2016-10-11 19:44:22 +02:00
Nikita Popov
a910f6a247 Remove $separator arg from Name::toString() 2016-10-09 12:45:15 +02:00
Nikita Popov
caa5c0cc76 Graceful handling for "special" errors
Nearly all special errors are now handled gracefully, i.e. the
parser will be able to continue after encountering them. In some
cases the associated error range has been improved using the new
end attribute stack.

To achieve this the error handling code has been moved out of the
node constructors and into special methods in the parser.
2016-10-09 12:38:18 +02:00
Nikita Popov
5e5cb86e83 Remove support for node cloning in traverser 2016-10-09 00:41:55 +02:00
Nikita Popov
f6eb341b15 Fix GroupUse prefix attribute assignment 2016-10-08 23:55:46 +02:00
Nikita Popov
2e5ae28c39 Don't use deprecated getMock() 2016-09-30 20:42:19 +02:00
Nikita Popov
c79ea6d1d3 Support recovery from lexer errors
Lexer::startLexing() no longer throws, instead errors can be fetched
using Lexer::getErrors().

Lexer errors now also contain full line and position information.
2016-09-30 20:23:36 +02:00
Nikita Popov
9e5d3bbe25 Remove Error::(get|set)RawLine()
These have been superseded by Error::(get|set)StartLine().
2016-09-30 18:30:01 +02:00
Nikita Popov
f3c7dc9d89 Add Error::getMessageWithColumnInfo() 2016-09-30 18:24:43 +02:00
Nikita Popov
c5e0c3d7e2 Catch lexer errors in throwOnError=0 mode 2016-09-30 13:49:34 +02:00
Nikita Popov
ea47b6e0d6 Add NodeAbstract::setDocComment() 2016-09-17 20:51:22 +02:00
Nikita Popov
9e1c535b1d Update builder tests to use "flags" instead of "type" 2016-09-17 20:51:22 +02:00
Nikita Popov
1ab24d26ee Add support for PHP 7.1 types to builders
This adds support for void, iterable and nullable types.
2016-09-16 13:53:13 +02:00
Nikita Popov
a7120116b0 Merge branch '2.x' 2016-08-30 22:38:34 +02:00
Nikita Popov
83f34e7fa4 Retain comments on blocks on first inner statement 2016-08-30 22:37:51 +02:00
Nikita Popov
d0cfb98133 Merge branch '2.x'
Conflicts:
	lib/PhpParser/Parser/Php7.php
2016-08-30 22:14:09 +02:00
Nikita Popov
13f7321def Forbid "=& new" in PHP 7 mode 2016-08-30 22:12:01 +02:00
Nikita Popov
f7cb00d6d3 Add missing canonicalization in test 2016-07-25 21:02:53 +02:00
Nikita Popov
7dae6c7a6b Implement JsonSerializable for Nodes and Comments
Exposes the properties and adds an additional nodeType property.
2016-07-25 20:59:09 +02:00
Nikita Popov
977cbab8e7 Decrement errorState when recovering from -> error
It's likely that an error after -> will trigger another one due to
missing semicolon without shifting a single token. We prevent an
immediate failure in this case by manually setting errorState to 2,
which will suppress the duplicate error message, but allow error
recovery.
2016-07-25 17:37:54 +02:00
Nikita Popov
09086fbe0a Support partial parsing of $foo->
Introduce Error node for this purpose.
2016-07-25 17:03:58 +02:00
Nikita Popov
ec614c95dd Add hasLeadingNewline attribute to InlineHTML
Use this attribute to not print an extra newline if the original
code did not have it.
2016-07-25 16:44:25 +02:00
Nikita Popov
b740076ab1 Remove deprecated Name::set*() methods 2016-07-25 14:50:37 +02:00
Nikita Popov
c9fea2ef67 Extend Name::slice() to support negative length+offset 2016-07-25 14:47:24 +02:00
Nikita Popov
5f97b12576 Introduce explicit Finally node 2016-07-25 14:25:04 +02:00
Nikita Popov
1dea9111a2 NodeDumper: Resolve type for include/use as well 2016-07-25 14:04:04 +02:00
Nikita Popov
174e6c3cab NodeDumper: Print modifiers as strings 2016-07-25 13:53:49 +02:00
Nikita Popov
18129480ae Rename $type subnode to $flags
Type makes it sound like a type-hint, and on a number of other nodes
$type is used for exactly that. Use $flags to hold modifiers instead.
2016-07-25 13:33:19 +02:00
Nikita Popov
867ae5148d Bring Trait constructor in line with Class/interface 2016-07-22 17:01:51 +02:00
Nikita Popov
537b59d4d1 PHP 7.1: Support multi-catch
Catch::$type is now an array Catch::$types.
2016-07-22 15:40:00 +02:00
Nikita Popov
7ff12b8fcb Remove deprecated Comment methods 2016-07-09 22:00:39 +02:00
Nikita Popov
574665b45b PHP 7.1: list() with keys
Expr\List will now contain ArrayItems instead of plain variables.
I'm reusing ArrayItem, because code handling list() must also handle
arrays, and this allows both to go through the same code path.

This also renames Expr\List->vars to ->items.

TODO: Should Expr\List be dropped in favor of Expr\Array with an
extra flag?
2016-07-09 21:55:55 +02:00
Nikita Popov
437890d386 PHP 7: Short destructuring syntax
Potentially the pretty printer should force use of [] in assignment
context, instead of relying on the existance of the right attribute.
2016-07-06 23:43:23 +02:00
Nikita Popov
1edf72c040 PHP 7: Support nullable types
Using a new NullableType node.
2016-07-06 18:36:18 +02:00
Nikita Popov
5ea2a76d80 PHP 7.1: Class constant visibility support 2016-07-05 23:01:06 +02:00
Nikita Popov
5044fce1ff PHP 7.1: Add void+iterable support
In PHP 7 mode, these will now be represented as strings 'void'
and 'iterable'.
2016-07-05 22:35:27 +02:00
Nikita Popov
96cbd48df6 Improve error recovery quality
In particular, support recovering from a missing trailing semicolon,
while keeping the rest of the expression.
2016-04-20 16:47:50 +02:00
Nikita Popov
e45e31c218 Fix new.test and code test runner
Using only the basename leads to collisions...
2016-04-20 15:03:18 +02:00
algo13
371c783344 Escape all low control characters in strings 2016-04-19 17:45:39 +02:00
Nikita Popov
5a6e7dd452 Avoid superfluous newline when printing Nops 2016-04-19 15:10:51 +02:00
Nikita Popov
90eb1165d1 Switch Exit kind attribute to flag
To be consistent with everything else.
2016-04-19 14:51:05 +02:00
Nikita Popov
e7869b9f14 Forbid invalid octals in PHP 7 mode 2016-04-18 13:59:18 +02:00
Nikita Popov
4c7ad7e194 Allow nop statements before namespace declaration 2016-04-16 21:39:49 +02:00
lvht
39f93f09f9 Add setReturnType() method to function/method builders
Also support scalar type hints in existing setTypeHint() method on params.
2016-04-09 18:41:38 +09:00
Nikita Popov
8cacc85913 "Fix" HHVM build 2016-04-07 12:27:08 +09:00
Nikita Popov
588e6a4d4c Add string kinds and doc string labels
Scalar\String_ and Scalar\Encapsed now have an additional "kind"
attribute, which may be one of:

 * String_::KIND_SINGLE_QUOTED
 * String_::KIND_DOUBLE_QUOTED
 * String_::KIND_NOWDOC
 * String_::KIND_HEREDOC

Additionally, if the string kind is one of the latter two, an
attribute "docLabel" is provided, which contains the doc string
label (STR in <<<STR) that was originally used.

The pretty printer will try to take the original kind of the string,
as well as the used doc string label into account.
2016-04-07 12:09:00 +09:00
Nikita Popov
52cb5ecec7 Fix comment reformatting for some cases
It did not correctly deal with non-standard comments containing
indented lines.
2016-04-07 11:47:40 +09:00
Nikita Popov
14de71898e Fix nop statement comment assignment
Keep around the start attributes on the lookahead token around in
a separate parser property.
2016-04-05 19:19:15 +09:00
Nikita Popov
5fa8493675 Make canonicalization less aggressive 2016-04-05 10:37:45 +09:00